Cloud Unit Economics
The math explaining why high cash burning cloud companies *might* make sense
The reason cloud companies can command high valuation multiples is because they can scale quickly and print money at scale. Investors and operators typically use revenue multiples as short hand for estimating the valuation of a company but many folks fail to understand the underlying unit economics of why those multiples make sense.
A company’s valuation is ultimately determined by a company’s future cash flow potential, which hinges on its unit economics. Understanding unit economics will drive better decisions for both operators and investors. Too many people become blinded by short-hand valuation multiples or metrics that they don’t even really understand what is driving long-term value.
Unit Economics Foundation
When people talk about “unit economics” in software they are typically referring to how adding an additional customer or $1 of revenue impacts the financials over the life of the customer.
Unit economics answers the following: How much money can a company make from selling its product to a customer.
Acquiring Customers
The issue with cloud companies is they have to front load so much cost to acquire a new customer that they lose money for a long time and only if the customer continues to renew does the company actually make money.
The Customer Acquisition Cost (CAC) Payback Period shows how long it takes for a company to recoup its customer acquisition costs (sales & marketing spend).
There are plenty of potential nuances with how you calculate the payback period, but that is not the purpose of this post. Just understand it is suppose to represent how long it takes to breakeven on the sales & marketing (S&M) costs it took to acquire a customer.
Below is a great chart from Lenny Rachitsky on the general rules-of-thumb for good payback periods.
Here are some things you should understand about these payback periods:
Payback period is not really an efficiency metric. Rather it’s a risk metric that shows how long it takes to get your money back (h/t Dave Kellogg for driving home this point). The unit economics can be terrible on a company with a 6 month payback period if customers typically churn in month 7.
Longer payback periods are acceptable as a company moves up market to the enterprise segment because those companies typically stay customers longer and have bigger wallets to expand.
Payment terms can significantly impact the cash impact of the payback period and the cash flow trough discussed below.
Payback period only shows time to breakeven on sales & marketing costs. It doesn’t speak to research & development (R&D) or general & administrative (G&A). For companies heavily investing in R&D it will take even longer to get to cash flow breakeven.
The Cash Flow Trough
The CAC payback period explains why fast growing SaaS companies are burning large amounts of cash and the cash flow trough illustrates how this impacts cash burn.
I first saw the visualization of the SaaS cash flow trough from David Skok. The faster a company is growing the larger the cash flow trough because of the individual customer CAC payback periods being stacked on top of each other faster than profits accumulate from customers on the other side of the cash trough.
But as growth rates start to slow those companies see faster and larger cash inflows. The trick is balancing the cash flow trough with a company’s cash runway and fundraising timing.
If the unit economics are great then companies will want to hit the gas and grab as many customers as possible. However, this means the cash flow trough gets deeper and the company will burn lots of money initially (in hopes that lots more money comes later). This means companies need to carefully manage their cash position and timing for raising more money.
The visualization above shows why good unit economics is critical. If the unit economics are terrible then you will never get out of the cash flow trough…
The Burn Multiple
There are more costs to recoup then just S&M. Using the example income statement below there are three things covered in CAC payback:
Gross profit, which is made up of:
Revenue
Cost of revenue
Sales & marketing (S&M)
But after S&M is recouped, the aggregate profits must also recoup costs from the below items before a company is cash flow positive:
Research & development (R&D)
General & administrative (G&A)
If the unit economics are bad, then you won’t generate enough cash flow from customers to recoup the investments in these areas.
This leads to another important metric (but not really indicative of good/bad unit economics): The Burn Multiple. The burn multiple shows how much cash a company burns to generate $1 of net new annual recurring revenue (ARR).
I like the burn multiple because you can’t easily manipulate it or hide anything since it is inclusive of all cash that is burned. Good/bad burn multiples are a function of 1) growth rates, 2) stage of company, and 3) company specific circumstances (like complexity of product) so these factors must be understood as well.
Below is a survey from ICONIQ portfolio companies and their burn multiples by ARR stage. As ARR increases the burn multiple should decrease because growth rates are slower (meaning more customers are exiting the cash flow trough and driving profits) and more operating leverage is being built.
While the burn multiple can be a great high-level overview of capital efficiency, it doesn’t tell you how good the unit economics are. Even if a company has a great burn multiple the business can still be garbage if the unit economics are bad.
Eventually the ARR per employee must surpass the expenses per employee for a company to be profitable…
The magic of SaaS is that if customers renew then the marginal cost in the outer years is relatively small and you therefore have incredible profits each additional year you keep customers, which is the real power of SaaS unit economics…
Unit Economic Metrics
Up until this point I have covered why SaaS companies lose money for so long and provided some metrics that may somewhat indirectly speak to unit economics.
Unit economics is like an income statement at the individual customer level. In theory the best way to understand the unit economics is to look at the following:
Lifetime value of a customer (LTV)
Customer acquisition Cost (CAC)
I have already defined CAC above. LTV is defined below and represents the total money brought in from the customer over their lifetime after reducing for the direct ongoing costs to manage the software (reflected below by gross margin %).
Adjusting LTV with gross margin % standardizes LTV amongst cloud companies that have a wide range of gross margin percentages — making comparisons easier.
You can now use LTV to tell you how much dollars a customer provides to pay down operating expenses: R&D, S&M, and G&A.
The direct and incremental costs associated with acquiring a customer is S&M (or CAC) while R&D and G&A should decrease as a % of revenue as companies build operating leverage. This is why a popular SaaS metric for unit economics is LTV/CAC
Common rule of thumb: A good LTV/CAC >= 3
First off, an LTV/CAC ratio of 3 is actually kind of a low bar. Most great companies are something closer to >6.
Example: A company sold a deal with an LTV of $120K and CAC of $40K — LTV/CAC of 3. This means there is $80K left over to pay for the R&D and G&A costs and then hopefully lots leftover for the company in actual profits.
Context is critical with all metrics. What makes a good LTV/CAC ratio also depends on the other expenses. For example, product-led growth (PLG) companies tend to spend a lot more on R&D. Theoretically, CAC could be low because a company is spending a crazy high amount in R&D. In this case a high LTV/CAC ratio is not as attractive because there is extra R&D costs to pay down before profits accrue to the company.
Expansion
Another aspect of a customer’s LTV is expansion opportunities that happen over the customer life. Companies that can expand a customer over time increase the actual LTV. This is typically not reflected in the LTV/CAC ratio so companies with higher expansion potential should be looked at more favorably. Expansion potential is reflected in a company’s net revenue retention (NRR) metric. NRR is similar to compound interest…
One of the reason expansion revenue is so great is that it has a lower cost than new business revenue so it provides more dollars to the bottom line. So businesses with more expansion opportunity will have much more favorable unit economics. Based on a KeyBanc survey, expansion dollars are about 1/3rd the cost of new business.
Word of Caution Using LTV
LTV:CAC is in theory a fantastic metric to understand unit economics but it comes with one major flaw…churn assumptions. To get to the customer life we use 1/churn rate. So if a company has an annual churn rate of 10% then there is an assumed 10 year customer life.
Very early stage companies have no idea what their actual customer life will be so anything is just a wild guess best on benchmarks. Also, a lot of companies are still using benchmarks and churn history from that last decade of boom times for the cloud. The future likely looks different — especially with increasing competition, generative AI, etc. Most companies should be more conservative with LTV calculations.
Strategic Advantage from Efficient CAC
Companies can gain strategic advantages from many things — technology, distribution, data, etc.
But an efficient go-to-market (GTM) motion can also be a huge strategic advantage. If a company can acquire customers significantly cheaper than its competition, then they can spend a lot more cash and accelerate past competitors. Unit economics shouldn’t just be something you think about when you are at significant scale.
Stacking S-Curves
The idea of “stacking S-curves” is to find the next lever for growth before the current one flattens and falls.
Companies in the process of stacking an S-curve will likely require a sacrifice for a time of the unit economics and cash burn. But if successful then it will generate higher and faster cash flows in the future. If unsuccessful though then the rest of the business with good unit economics will subsidize the bad one until it improves or is killed. This is why unit economics should be segmented.
Final Thoughts
Companies are ultimately valued by their future cash flows. Understanding a company’s unit economics is critical in determining how profitable a company can eventually become at scale.
The cloud dream is that at scale these companies can have 30%+ free cash flow margins and then sustain that free cash flow level over a long period of time. This is why SaaS companies can have very high valuations relative to their current revenue levels — fast growth combined with incredibly high future cash flows.
But most SaaS companies won’t get to this cash flow wonderland, which is OK, but it means the company should be less valuable.
Understanding unit economics can help a company understand how profitable it can become and it also helps inform important decisions to drive more growth/profits:
Helping discover bad investments
Guide pricing strategy / optimizations
Discover high ROI sales channels
Determine best performing customer segments, products, geography, etc.
CEO and operators must understand unit economics so they can make better long-term decisions based on what matters - maximizing future cash flow potential.
Pro Tip: If you are trying to explain to a family member why it makes sense for software companies to lose money for so long then you need to understand SaaS unit economics :). Share this with them!
CLOUD COMPUTING UNIT ECONOMICS
A key issue here is that Cloud computing is incredibly sticky.
Take the market leader, AWS as an example.
Companies, particularly the large enterprises, use so much data that it would have taken years and posed security challenges to upload it over the internet. So Amazon made it easy for customers to switch to AWS through the introduction of a 100 Terabyte super secure data device, which they called a Snowball, that they would send to a company’s office. The company would transfer data from its on premises infrastructure to the Snowball, which was then sent to Amazon to be uploaded to the cloud infrastructure. Problem solved. As computing power and data consumption increased, Amazon offered what they called a Snowmobile, which was a huge truck full of Snowballs that would be used to transfer the enormous quantities of data to the Amazon cloud. Even then it could take months to complete the process.
To give you an idea of the scale of the process, Amazon transferred its own data from Oracle infrastructure to its AWS cloud and the process took 13 years from start to finish, commencing in 2006 and completing in 2019.
So once this process is complete, people need a very good reason to switch to another provider.
It is a cliche, but the value of a business is the discounted value of future cash flows. It isn't about what you earn today, it's about optimizing what you will earn tomorrow. Every new customer increases tomorrows cash flows exponentially and this compounds over time. NPV that back to today and you have a valuable company that is growing in value all the time, regardless of today's earnings numbers.